OpenAI: o3 vs OpenAI: o4 Mini
OpenAI: o4 Mini is cheaper than OpenAI: o3 by about 45% on a blended (75% input / 25% output) basis. OpenAI: o3 charges $2.000000 / $8.000000 per 1M input/output tokens; OpenAI: o4 Mini charges $1.100000 / $4.400000.
| Field | OpenAI: o3 | OpenAI: o4 Mini |
|---|---|---|
| Input $/Mtok | $2.000000 | $1.100000 |
| Output $/Mtok | $8.000000 | $4.400000 |
| Cache read $/Mtok | $0.500000 | $0.275000 |
| Cache write $/Mtok | – | – |
| Input $/Mchar | 0.563428 | 0.309885 |
| Output $/Mchar | 2.253712 | 1.239541 |
| Context window | 200000 | 200000 |
| Tokenizer | o200k | o200k |
Price history
Same workload, monthly cost
Illustrative monthly cost at 10M tokens/month, split per workload preset. Not a substitute for the calculator, which uses your own volume, cache and batch assumptions.
| Workload | OpenAI: o3 | OpenAI: o4 Mini |
|---|---|---|
| Coding agent | $23.00 | $12.65 |
| RAG chat | $29.00 | $15.95 |
| Summarisation (batch) | $26.00 | $14.30 |
| Classification | $21.20 | $11.66 |
| Chat assistant | $44.00 | $24.20 |
| Content generation | $68.00 | $37.40 |
Cite this page
OpenAI: o3 vs OpenAI: o4 Mini: $2.000000/$8.000000 vs $1.100000/$4.400000 per 1M tokens, as of 2026-10-04 (Token Barometer, https://tokenbarometer.com/compare/openai_o3-vs-openai_o4-mini).